Re: [cross-project-issues-dev] Improving Hudson (was Why allowing Hudson to write to your downloads is a Bad Idea.)
Pascal Rapicault <pascal@xxxxxxxxxxxx>
Cross project issues
09/14/2011 01:32 PM
Improving Hudson (was Why allowing Hudson
to write to your downloads is a Bad Idea.)
Unfortunately it is not only argument sake... I think
everybody understands the possibilities/benefits of a CI system because
they've experienced it at their companies but the one available at Eclipse
is hard to work with because of the restrictions imposed. Your initial
suggestion of reducing the ability to move files around, which IMO is fundamental
to a build system, is just causing all these moments of pain dealing with
hudson at eclipse to resurface.
At a time where the foundation is thinking of providing
a common build infrastructure in the context of the LTS initiative, it
seems that this decisions / recommendations are being at odds with the
overall goal of making it easy to build at Eclipse (and fuel the discussions
of continuing building things being corporate firewalls: the infrastructure
is hard to work with and can't be trusted (this is the conclusion someone
would get looking at this thread)).
At this point I don't have a particular solution to offer
for the given problem and instead I'm proposing to start list of the "freedoms"
committers need to have to be able to successfully leverage the infrastructure
put in place. So here it is:
- Need to request creation of a build job.
you trust me to write code and commit it but not to create
- Difficulties to follow the typical workflows of deploying
from Hudson (be it a file system, or a maven repo)
Btw, the cron approach may ends up partially copying repos.
Also it does not anyone to control the complete end to end build from Hudson.
- Inability to get a stack dump from a slave when the
test are the build is stuck
On 2011-09-14, at 11:27 AM, Denis Roy wrote:
I never said anyone should do any of that. I never said
you shouldn't trust the output of Hudson either.
I only suggested what is in the subject line of this email: allowing Hudson
to write to your downloads is a Bad Idea. And I explained why.
We simply got here for argument's sake.
On 09/14/2011 11:15 AM, Chris Recoskie wrote:
So... suddenly our builds will take days for someone to
build it, grab it, test it, and promote it, and even then it's highly unlikely
that they'd catch anything that a somewhat competent attacker would throw
at us. There's no way anyone is going to take on this overhead.
Team Lead, IBM CDT and RDT
<unnamed.gif>Denis Roy ---09/14/2011
10:09:39 AM---On 09/14/2011 10:02 AM, Ed Merks wrote: > I agree with
On 09/14/2011 10:02 AM, Ed Merks wrote:
I agree with Doug.
At no point have I seen anyone answer this question:
What can be done manually to determine if what's produced
by Hudson is compromised or not?
Off the top of my head:
1. run a build on a remote system and compare the pre-signed binaries.
2. run a past build and compare today's binaries with those in the past.
3. run a build and examine the execution trace.
4. run a build, run the executable and examine network output for unknown
I also have to question whether this change during the
SR1 shutdown phase is appropriate timing...
Go download the latest Linux Kernel from Kernel.org
and tell me if there is ever a more appropriate time than 'now' to discuss
On 14/09/2011 6:55 AM, Schaefer, Doug wrote:
I'll come back to something Dave Carver mentioned
yesterday. If we don't trust Hudson, then we shouldn't be using it, or
at least should be wrapping it up in tighter security, like a VPN for example.
If someone is going to do something malicious and they're smart, you're
likely not going to be able to discover it. You have to cut it at the source.
And is this not an issue other Hudson/Jenkins users have run into? What
are they doing for security. Or do they trust Hudson as much as they do
cross-project-issues-dev mailing list
cross-project-issues-dev mailing list